CHAPTER 3 EXPERIMENTAL STUDY OF INTERFACE BEHAVIOR BETWEEN COMPOSITE PILES AND TWO SANDS

Transcription

1 CHAPTER 3 EXPERIMENTAL STUDY OF INTERFACE BEHAVIOR BETWEEN COMPOSITE PILES AND TWO SANDS 3.1 INTRODUCTION The properties of materials used in composite piles can differ significantly from those of conventional pile materials, e.g., steel and concrete, including differences in anisotropy, stiffness, surface hardness, and surface roughness. These differences may produce differences in the behavior of interfaces between the piles and the surrounding soil. Interface behavior plays an important role in the ultimate capacity and load transfer characteristics of the pile, as shown schematically in Figure 3.1. For example, the axial pile capacities of a composite pile and a conventional pile, with identical dimensions and soil conditions, may be quite different due to differences in the developed shaft capacity. To date, most laboratory studies published in this area involve interface shear tests on traditional pile materials. Relatively few studies are available on interface shear behavior between composite pile materials and soils. The principal focus of this chapter is on the soil-pile interface behavior characteristics of composite piles against sand. This chapter summarizes interface shear tests results for seven pile material types (5 composite pile materials, and 2 conventional pile materials) against two types of sands. The interface behavior of the different piles is studied within the geotribology framework proposed by Dove and Jarett (22). 28

2 This chapter is divided into five sections: 1. Soil materials. Describes the gradation and strength properties of each of the soils used for interface testing 2. Pile surfaces. Describes the surface texture and hardness of the different interface materials tested. 3. Interface shear tests. Describes the equipment and procedures used for interface shear testing, and summarizes the results obtained. 4. Discussion of results. 5. Summary. Pile capacity Q ult = Q s + Q p τ Soil-pile interface shear tests: σ n2 σ n3 σ n1 Shaft capacity Q s = f (δ) τ Interface friction angle, δ δ peak δ cv Tip capacity Q p σ n1 σ n2 σ n3 σ n Figure 3.1. Influence of soil-pile interface friction on pile capacity A test matrix summarizing the tests carried out for the interface behavior study is provided in Table

3 Table 3.1 Interface behavior test matrix Type of Test Type of Soil Type of Pile Surface Comments Direct shear test Density sand Not applicable Average D r = 7 % Direct shear test Density sand Not applicable Average D r = 1 % Direct shear test Model sand Not applicable Average D r = 75 % Interface shear test Density sand Lancaster CP4 Curved specimen, 24-in. nominal diameter Interface shear test Model sand Lancaster CP4 Curved specimen, 24-in. nominal diameter Interface shear test Density sand Hardcore 24-4 Curved specimen, 24-in. nominal diameter Interface shear test Model sand Hardcore 24-4 Curved specimen, 24-in. nominal diameter Interface shear test Density sand Hardcore FRP, Flat plate, surface untreated no bonded sand with no acrylic coating Interface shear test Model sand Hardcore FRP, Flat plate, surface untreated no bonded sand with no acrylic coating Interface shear test Density sand Hardcore FRP, Flat plate, textured surface with bonded sand with bonded sand Interface shear test Model sand Hardcore FRP, Flat plate, textured surface with bonded sand with bonded sand Interface shear test Density sand Plastic Piling Flat coupons from prismatic recycled plastic rectangular beams Interface shear test Model sand Plastic Piling Flat coupons from prismatic recycled plastic rectangular beams Interface shear test Density sand Concrete Coupon from prestressed concrete pile Interface shear test Model sand Concrete Coupon from prestressed concrete pile Interface shear test Density sand Steel Coupon from sheet pile Interface shear test Model sand Steel Coupon from sheet pile 3

4 3.2 SOIL MATERIALS Two sands were used in this study: Density sand and a Model sand. The Density sand is a fine to medium grained, silica sand with subrounded to rounded grains, and the Model sand consists of a blend of Light Castle sand with Bedding sand. The Model sand consists predominantly of fine grained sand with subangular to angular grains. The Density sand was used in a previous study by Gómez et al. (2). The Model sand was prepared to mimic the foundation soils of the pile test site at the Route 351 Bridge Index properties Index parameter values for both sands used in this study are listed in Table 3.2. Grain size distribution curves for these sands are shown in Figure 3.2. The Density sand has grain sizes ranging from.2 to.9 mm with no fines. The Model sand has grain sizes ranging from.1 to 2 mm with about 5 percent fines. Microscopic views of the sands are shown in Figure 3.3. Table 3.2. Index parameter values of the sands used in this study Parameter Density Sand (1) Model Sand ASTM Standard D 1 (mm).3.8 D 5 (mm).5.18 C u D2487 C c γ max (kn/m 3 ) D4253 γ min (kn/m 3 ) D4254 G s D854 Notes: 1) After Gómez, Filz, and Ebeling (2) 31

5 GRAVEL Coarse to medium SAND Fine SILT CLAY Percentage Finer Density Sand Model Sand Particle diameter (mm).1.1 Figure 3.2. Grain size curves of test sands.5 mm.5 mm a) Density sand b) Model sand Figure 3.3 Microscopic views of the test sands 32

6 3.2.2 Direct Shear Tests of Sands Direct shear tests were performed to determine the internal friction angles of the Density and Model sands. Displacement controlled direct shear tests with a 11.6 mm (4-inch) square shear box were performed. All tests were carried out at a horizontal displacement rate of.9 mm/min (.36 in/min). The direct shear tests were performed in accordance with ASTM Standard D38-98 (ASTM 1998). Samples were prepared by means of air pluviation and tamping. Dry sand was poured through a 6.5 mm funnel positioned at 1 mm above the top of sand surface. Increased density was achieved by means of a small tamping device. To minimize the friction between the sand and the soil box, the inside of the walls were coated with a thin film of vacuum grease. Direct shear test results, in terms of shear stress versus horizontal displacement curves, are presented in Figures 3.4 through 3.6 for the medium dense to dense Density and Model sands, respectively. For the entire range of relative densities and normal pressures tested, all of the direct shear tests exhibited a peak shear stress followed by a stress reduction towards a residual shear stress with a nearly constant volume state. The peak and residual (nearly constant volume state) friction angles obtained from the direct shear tests are summarized in Table 3.3. Table 3.3. Internal friction angles obtained from direct shear tests (1) Parameter Density Sand (Average D r = 7 %) Density Sand (Average D r = 1 %) Model Sand (Average D r = 75 %) Peak φ p 34.7 o 39.3 o 43.4 o Residual φ cv 29.6 o 29 o 36.2 o Notes: 1) Four normal stresses were used: 25, 5, 1 and 2 kpa 33

7 Initial σn = 22.3 kpa Shear Stress (kpa) Initial σn = 11.6 kpa Initial σn = 51.4 kpa 25 Initial σn = 23.7 kpa Horizontal Displacement (mm) a) Shear stress displacement curves 15 φ' peak = 34.7 o Shear stress (kpa) 1 5 Non-linear envelope: φ ο = o φ = 6.2 o φ' cv = 29.8 o b) Shear strength envelopes Normal stress (kpa) Figure 3.4 Direct shear test results for Density Sand (Average D r = 7 %) 34

8 Shear Stress (kpa) Initial σn = 22.3 kpa Initial σn = 11.6 kpa Initial σn = 51.4 kpa 25 Initial σn = 24.8 kpa Horizontal Displacement (mm) a) Shear stress displacement curves Shear stress (kpa) Non-linear envelope: φ ο = 41.5 o φ = 4.6 o φ' peak = 39.3 o φ' cv = 29 o b) Shear strength envelopes Normal stress (kpa) Figure 3.5 Direct shear test results for Density Sand (Average D r = 1 %) 35

9 Initial σ n = kpa Shear Stress (kpa) Initial σ n = 11.8 kpa Initial σ n = 51.5 kpa 25 Initial σ n = 23.7 kpa Horizontal Displacement (mm) a) Shear stress displacement curves 15 φ' peak = 43.4 o Shear stress (kpa) 1 5 Non-linear envelope: φ ο = 44.6 o φ = 9.8 o φ' cv = 36.2 o Normal stress (kpa) b) Shear strength envelopes Figure 3.6 Direct shear test results for Model Sand (Average D r = 75 %) 36

10 D r, in Table 3.3, refers to the sample relative density which can be calculated as follows: where: D r 1 1 γ min γdry = 1% 1 1 γ γ min max (3.1) γ max and γ min = are the maximum and minimum dry unit weights for each sand (values given in Table 3.2), γ dry = dry unit weight of the samples in the direct shear tests. Both linear and curved peak shear strength envelopes are shown in Figures 3.4 through 3.6. The non-linear shear strength envelopes were obtained using the following expression proposed by Duncan et al. (198): φ φ φ σ n = o log1 P a (3.2) where: φ o = peak secant friction angle at a normal stress equal to the atmospheric pressure (P a = 11.4 kpa), φ = reduction in peak secant friction angle for a tenfold increase in normal stress, σ n. The values of φ o and φ for both sands are shown in Figures 3.4 through 3.6. The shearing area of the soil specimens decreases during direct shear testing. The direct shear test stress paths, shown in Figures 3.4 through 3.6 in τ-σ space, have been corrected for this effect. 37

11 3.3 PILE SURFACES Introduction The interface testing program included determining interface friction angles between the two sands described above and the following seven pile surfaces: Lancaster FRP shell (curved), Hardcore FRP shell (curved), Hardcore FRP plate without bonded sand (flat), Hardcore FRP plate with bonded sand (flat), Plastic Piling recycled plastic (flat), steel sheet pile (flat), and prestressed concrete (flat). The following subsections describe the above pile surfaces within a geotribology framework Surface topography characterization The importance of surface topography and the degree of surface roughness in interface behavior and soil-structure interaction response has been noted by many investigators (e.g. Yoshimi and Kishidi 1982, Kishida and Uesegui 1987, Lehane et al. 1993, Dove et al. 1997, Han 1997, Dove and Frost 1999, Dove and Jarrett 22). The surface roughness of each pile material investigated was measured with a Taylor - Hobson Talysurf stylus profilometer. A schematic of the profilometer used in this study is shown in Figure 3.7. As the figure illustrates, the profilometer comes equipped with a diamond stylus tip that traverses the surface to be characterized. Transducers in the profilometer help record the vertical and horizontal coordinates of the stylus tip. Therefore, the raw surface profile data consists of a series of discrete coordinate pairs (x, z) that can be postprocessed and analyzed to obtain the surface topography characterization of the material investigated (Dove and Jarett 22). Surface topography was quantified by three commonly used roughness parameters: the maximum peak to valley height, R t ; the average mean line spacing, S m ; and the arithmetic average roughness, R a. S m is twice the mean distance between locations at which the profiles crosses the centerline drawn through the centroid of the profile. R a is the 38

12 arithmetic average value of the profile departure from the mean line along the profile length. These three roughness parameters are illustrated in Figure 3.8. More detailed discussion of these parameters is given in ISO Standard 4287 (ISO 1997). Figure 3.7 Stylus profilometer sketch (Johnson 2) z Max. peak R t Mean Line x Max. valley S 1 S 2 S 3 S 4 S i S m 1 i = n = n i= 1 S i L 1 Ra = zx ( ) dx L Figure 3.8 Graphical representation of roughness parameters R t, S m, and R a 39

13 Coupons approximately 25 mm x 2 mm were cut from each pile material. A series of six profiles, each 49 mm in length were made on the surface of each coupon. All profiles were made in the longitudinal direction of the piles. Photos of the seven pile surfaces and corresponding representative surface profiles are shown on Figures 3.9 to The roughness parameters for the seven interface materials are summarized in Table 3.4. The parameter values reported and used subsequently are averages of the six individual values obtained for each coupon. 5 mm (mm) a) Surface texture (mm) b) Surface roughness profile Figure 3.9 Surface characteristics of Lancaster FRP composite pile 4

14 5 mm (mm) (mm) a) Surface texture b) Surface roughness profile Figure 3.1 Surface characteristics of Hardcore FRP composite pile 5 mm (mm) (mm) a) Surface texture b) Surface roughness profile Figure 3.11 Surface characteristics of Hardcore FRP plate 41

15 5 mm (mm) (mm) a) Surface texture b) Surface roughness profile Figure 3.12 Surface characteristics of Hardcore surface treated FRP plate 5 mm (mm) (mm) a) Surface texture b) Surface roughness profile Figure 3.13 Surface characteristics of Plastic piling plastic composite pile 42

16 5 mm (mm) (mm) a) Surface texture b) Surface roughness profile Figure 3.14 Surface characteristics of prestressed concrete pile 5 mm (mm) a) Surface texture (mm) b) Surface roughness profile Figure 3.15 Surface characteristics of steel sheet pile 43

17 Table 3.4. Summary of surface roughness measurements Roughness Parameters Interface Material Type R t (µm) S m (mm) R a (µm) Average S. D. Average S. D. Average S. D. LC - CP HC HC untreated FRP plate HC treated FRP plate PPI plastic Concrete Steel Notes: LC = Lancaster Composites Inc., HC = Hardcore Composites Inc., PPI = Plastic Pilings inc., R t = Maximum peak to valley height, S m = Average mean line spacing, R a = arithmetic average roughness, S.D. =Standard deviation Interface hardness The hardness of the interface material is another important factor for consideration when studying interface behavior. As the surface hardness of the construction material decreases, it becomes possible for the soil grains to penetrate into the construction material and shear failure may occur by plowing of the soil grains along the material surface (Pando et al. 22). In their study of sand-polymer interfaces, O Rourke et al. (199) found that the interface frictional strength decreased with increasing polymer hardness. The authors found that relatively hard and smooth polymer surfaces did not show significant dilatancy and the sand grains tended to slide along the polymer interface; whereas, the particle movement on relatively soft and smooth polymer surfaces involved predominantly rolling and temporary grain indentations that resulted in higher interface friction values. The findings by O Rourke et al. (199) are mainly applicable to smooth polymeric materials since, as the authors pointed out, surface roughness of rough polymers will promote dilation and result in increased shear resistance. The surface hardness of the different pile/interface materials was investigated by means of the Vickers hardness test. The Vickers hardness test has the flexibility to allow application of small loads, which was convenient for testing the softer materials such as recycled plastic and FRP composites, and consists of pressing a standard diamond pyramid into the sample. The Vickers hardness number (HV) is related to the load 44

18 applied and the area of the diamond pyramid indentation. The Vickers hardness tests were performed in accordance with ASTM Standard E (ASTM 1999). The results of the hardness tests are summarized in Table 3.5. Table 3.5. Surface hardness Interface Material Vickers Hardness (HV) Mean S. D. N LC - CP HC HC untreated FRP flat plate HC treated FRP flat plate Not measurable (1) PPI plastic Concrete Steel Notes: (1): Not measurable because it is difficult to clearly measure the diamond indentation mark due to presence of the bonded sand particles. HV = Vickers Hardness Number (ASTM Standard E384-99) (carried out at room temperature). LC = Lancaster Composites Inc., HC = Hardcore Composites Inc., PPI = Plastic Pilings inc., S.D. = Standard deviation, N = number of tests performed. 3.4 INTERFACE SHEAR TESTS Representative coupons were obtained from each of the seven different pile materials investigated. Careful consideration was given to selecting pile coupons that were representative of the average surface texture of the piles. The coupons from the plastic, concrete, and steel piles, and from the two Hardcore FRP plates, were all flat surfaced. Therefore, for these materials it was possible to use a flat shear box for interface shear testing. The coupons from the cylindrical FRP shells (i.e., Lancaster CP4 and Hardcore 24-4) were cut with a water-cooled saw to dimensions that fit into the bottom half of the shear box. However, because both composite FRP shells were fabricated with thermoset resins, their curvature could not be removed. To permit testing of these curved pile specimens, the top half of the shear box was modified such that it could conform to the outside curvature of the piles, as shown in Figure The interface shear tests carried out on the other pile materials involved the use of the unmodified top half of the direct 45

19 shear box. The interface shear tests were performed in accordance with ASTM Standard D (ASTM, 1997), and the horizontal displacement rate was.9 mm/min. Normal load Normal load Shear load SOIL Curvature in shear box to conform to pile SOIL Rigid substratum Side View FRP shell End View Figure 3.16 Sketch of modified interface shear test setup The interface tests performed in this study with the Density and Model sands are summarized in Tables 3.6 and 3.7, respectively. Interface shear tests were carried out at constant normal stresses ranging from 23 to 2 kpa (the contact area of the soil sample remained constant throughout the test). Typical interface shear stress versus interface displacement curves for the seven pile material types are shown in Figures 3.17 and 3.18 for the Density and Model sands, respectively. Additional details of the interface tests performed for this investigation are presented in Appendix A. Interface strength envelopes for the seven pile materials and for the Density and Model sands are shown in Figures 3.19 through Curvature of the strength envelopes with increasing normal stresses was generally not significant. The interface friction angles obtained from linear fits to the test data are summarized in Table

20 Table 3.6 Summary of interface shear test results on Density Sand Pile/Material D r (%) σ' n (kpa) S m (mm) R t (µm) Hardness (HV) S m /D 5 R t /D 5 τ p (kpa) d p (mm) τ cv (kpa) d cv (mm) Lancaster Composite FRP shell (CP4) Hardcore Composite FRP shell (24-4) Hardcore FRP plate untreated surface Hardcore See note (1) FRP plate See note (1) treated See note (1) surface See note (1) PPI plastic Prestressed concrete pile Steel from sheet pile Notes: (1) Not measurable because it is difficult to clearly measure the diamond indentation mark due to presence of the bonded sand particles. D 5 =.5mm for Density sand, D r =Relative density after sample consolidation (as per Eq. 3.1), σ n = Normal pressure, S m = Average mean line spacing, R t = Maximum peak to valley height, Hardness (HV) = Average Vickers hardness number, τ p =Peak interface shear stress, d p = Horizontal displacement at peak, τ cv = Residual interface shear stress, d CV = Horizontal displacement at residual. 47

21 Table 3.7 Summary of interface shear test results on Model Sand Pile/Material D r (%) σ' n (kpa) S m (mm) R t (µm) Hardness (HV) S m /D 5 R t /D 5 τ p (kpa) d p (mm) τ cv (kpa) d cv (mm) Lancaster Composite FRP shell (CP4) Hardcore Composite FRP shell (24-4) Hardcore FRP plate untreated surface Hardcore See note (1) FRP plate See note (1) treated See note (1) surface See note (1) PPI plastic Prestressed concrete pile Steel from sheet pile Notes: (1) Not measurable because it is difficult to clearly measure the diamond indentation mark due to presence of the bonded sand particles. D 5 =.18mm for Model sand, D r =Relative density after sample consolidation (as per Eq. 3.1), σ n =Normal pressure, S m = Average mean line spacing, R t = Maximum peak to valley height, Hardness (HV) = Average Vickers hardness number, τ p =Peak interface shear stress, d p = Horizontal displacement at peak, τ cv = Residual interface shear stress, d CV = Horizontal displacement at residual. 48

22 9 FRP plate with bonded sand σ' n = 17.2 kpa, D r = 63.7 % Interface shear stress, τ (kpa) 6 3 FRP plate - no treatment σ' n = 17.3 kpa, D r = 63.4 % PPI plastic σ' n = 15.4 kpa, D r = 6.8 % Hardcore 24-4 FRP shell σ' n = 1 kpa, D r = 66.3 % Lancaster CP4 FRP shell σ' n = 1 kpa, D r = 63.2 % Interface displacement (mm) a) Composite pile materials 9 Interface shear stress, τ (kpa) 6 3 Prestressed concrete pile σ' n = 11.9 kpa, D r = 63.3 % Steel sheet pile σ' n = 11.9 kpa, D r = 64.1 % Interface displacement (mm) b) Conventional pile materials Figure 3.17 Typical interface shear test results for Density Sand (σ n 1 kpa) 49

23 9 FRP plate - no treatment σ' n = 17.2 kpa, D r = 64.7 % FRP plate with bonded sand σ' n = 17.3 kpa, D r = 6.1 % Interface shear stress, τ (kpa) 6 3 PPI plastic σ' n = 17.3 kpa, D r = 65.1 % Lancaster CP4 FRP shell σ' n = 1 kpa, D r = 63.2 % Hardcore 24-4 FRP shell σ' n = 1 kpa, D r = 6.9 % Interface displacement (mm) a) Composite pile materials 9 Prestressed concrete pile σ' n = 11.9 kpa, D r = 61.9 % Interface shear stress, τ (kpa) 6 3 Steel sheet pile σ' n = 11.9 kpa, D r = 64.3 % Interface displacement (mm) b) Conventional pile materials Figure 3.18 Typical interface shear test results for Model Sand (σ n 1 kpa) 5

24 Interface shear stress (kpa) : Peak envelope : Constant volume (residual) envelope Model Sand δ p = 27.3 o, δ cv = 26 o Effective normal stress (kpa) Density Sand δ p = 19.7 o, δ cv = 16.6 o Figure 3.19 Interface shear strength envelopes for Lancaster Composite FRP shell Interface shear stress (kpa) : Peak envelope : Constant volume (residual) envelope Model Sand δ p = 29.5 o, δ cv = 29.3 o Density Sand δ p = 29.2 o, δ cv = 27.3 o Effective normal stress (kpa) Figure 3. 2 Interface shear strength envelopes for Hardcore Composite FRP shell 51

25 Interface shear stress (kpa) : Peak envelope : Constant volume (residual) envelope Model Sand δ p = 31.7 o, δ cv = 28 o Density Sand δ p = 28.4 o, δ cv = 25.7 o Effective normal stress (kpa) Figure 3.21 Interface shear strength envelopes for untreated Hardcore FRP plate 15 : Peak envelope : Constant volume (residual) envelope Interface shear stress (kpa) 1 5 Model Sand δ p = 37.3 o, δ cv = 32.6 o Density Sand δ p = 31.9 o, δ cv = 27.8 o Effective normal stress (kpa) Figure 3.22 Interface shear strength envelopes for treated Hardcore FRP plate 52

26 Interface shear stress (kpa) : Peak envelope : Constant volume (residual) envelope Model Sand δ p = 33.4 o, δ cv = 28.8 o Density Sand δ p = 27.6 o, δ cv = 24.9 o Effective normal stress (kpa) Figure 3.23 Interface shear strength envelopes for PPI plastic Interface shear stress (kpa) : Peak envelope : Constant volume (residual) envelope Model Sand δ p = 34.3 o, δ cv = 28. o Density Sand δ p = 33 o, δ cv = 27.7 o Effective normal stress (kpa) Figure Interface shear strength envelopes for concrete 53

27 15 : Peak envelope : Constant volume (residual) envelope Interface shear stress (kpa) 1 5 Model Sand δ p = 31.2 o, δ cv = 28.6 o Density Sand δ p = 28.2 o, δ cv = 25.1 o Effective normal stress (kpa) Figure Interface shear strength envelopes for steel Table 3.8 Summary of interface friction angles Soil Density Sand Model Sand Pile Soil properties Average D r (%) D 5 (mm) Surface Hardness Relative Roughness (HV) R t /D 5 S m /D 5 Interface Angles δ p δ cv ( o ) ( o ) LC FRP HC FRP Untreated FRP plate Treated FRP plate N/T PPI plastic Concrete Steel LC FRP HC FRP Untreated FRP plate Treated FRP plate N/T PPI plastic Concrete Steel Note: N/T = not testable. 54

28 3.5 DISCUSSION OF RESULTS It has previously been demonstrated that the strength and volume change behaviors of an interface system composed of a granular soil and a construction material are largely controlled by three factors: the relative size of the sand grains with respect to the surface asperity height and spacing, the hardness of the materials, and the soil grain shape (Kishida and Uesugi 1987, O'Rourke et al. 199, Irsyam and Hryciw 1991, Hryciw and Irsyam 1993, Dove and Harpring 1999, Dove and Frost 1999, Frost and Han 1999, and Dove and Jarrett 22). The interface shear test program conducted in this study was not designed to systematically investigate the different factors that influence the interface shear behavior of sands and FRP materials. Nevertheless, the interface friction angle values obtained in this study, summarized in Table 3.8, were obtained from interface shear tests having important differences in the following basic variables: 1. Mean grain size of sand, D 5 2. Angularity of sand 3. Maximum peak to valley height of pile surface, R t 4. Average mean line spacing of pile surface, S m 5. Pile surface Vickers hardness, HV The first two variables are related to soil characteristics, and the remaining three are related to pile surface characteristics. All of these variables, except angularity, were measured quantitatively. Therefore, the influence of the quantitative variables on the interface friction coefficients (peak and residual tan δ) can be investigated by performing multiple linear regressions. Linear regression analyses for both sand types and for peak and residual interface friction coefficients are presented in the following subsections. The regressions were carried out with the interface friction coefficient as the dependent variable, and R t /D 5, S m /D 5, and HV as independent variables. Due to small sample size (N=7), statistically significant results (at the 5% level) were not expected. The linear regression analysis results for both sand types and for peak and residual interface friction coefficients are presented in Sections through The general discussion of the regression results is provided in Section

29 3.5.1 Multiple linear regression for Density sand tan δ peak values Results of the regression analysis for the tan δ peak values from tests using the Density sand are summarized in Table 3.9. Table 3.9 Multiple linear regression result for tan δ peak of Density sand Regression equation R R 2 change due to R t /D R 2 change due to S m /D 5.32 R 2 change due to HV.5 R S tan δ = HV peak t 3 m 4 D5 D5 The coefficient of determination, R 2, for this fit was.66, shown graphically in Figure The contributions of R t /D 5, S m /D 5, and HV to R 2 were 53.6%, 45.6%, and.8%, respectively. This indicates that hardness, HV, is the least significant predictor of the tan δ peak values compared to R t /D 5 and S m /D 5. ( ).8.7 Rt 3 Sm 4 tan δ peak = HV D D (R 2 =.662, N = 7) 5 5 ( ) measured tan δ peak tan δ peak from multiple linear regression Figure 3.26 Goodness of fit for multiple linear regression on Density sand tan δ peak values 56

30 3.5.2 Multiple linear regression for Density sand tan δ cv values Results of the regression analysis for the tan δ cv values from tests using the Density sand are summarized in Table 3.1. Table 3.1 Multiple linear regression result for tan δ cv of Density sand Regression equation R R 2 change due to R t /D R 2 change due to S m /D R 2 change due to HV.4 R S tan δ = HV cv 2 t 2 m 4 D5 D5 ( ) The coefficient of determination, R 2, for this fit was.58, and is shown graphically in Figure The contributions of R t /D 5, S m /D 5, and HV to R 2 were 45.4%, 53.9%, and.7%, respectively. As in the previous analyses, this indicates that the hardness, HV, has considerably less influence on the tan δ cv values than R t /D 5 and S m /D Rt 2 Sm 4 tan δ cv = HV D D (R 2 =.581, N = 7) 5 5 ( ) measured tan δ cv tan δ cv from multiple linear regression Figure 3.27 Goodness of fit for multiple linear regression on Density sand tan δ cv values 57

31 3.5.3 Multiple linear regression for Model sand tan δ peak values Results of the regression analysis for the tan δ peak values from tests using the Model sand are summarized in Table Table 3.11 Multiple linear regression result for tan δ peak of Model sand Regression equation: R R 2 change due to R t /D R 2 change due to S m /D R 2 change due to HV.1 R S tan δ = HV peak 2 t 3 m 5 D5 D5 The coefficient of determination, R 2, for this fit was.89, and is shown graphically in Figure The contributions of R t /D 5, S m /D 5, and HV to R 2 were 68.6%, 31.3%, and.1%, respectively. As in the previous analysis, this indicates that hardness HV has considerably less influence on the tan δ peak values than R t /D 5 and S m /D 5. ( ).8.7 measured tan δ peak Rt 3 Sm 5 tan δ peak = HV D D 5 5 (R 2 =.892, N = 7) ( ) tan δ peak from multiple linear regression Figure 3.28 Goodness of fit for multiple linear regression on Model sand tan δ peak values 58

32 3.5.4 Multiple linear regression for Model sand tan δ cv values Results of the regression analysis the tan δ cv values from tests using the Model sand are summarized in Table Table 3.12 Multiple linear regression result for tan δ cv of Model sand Regression equation R R 2 change due to R t /D R 2 change due to S m /D R 2 change due to HV.2 R S tan δ = HV cv 2 t 3 m 4 D5 D5 The coefficient of determination, R 2, for this fit was.47, and is shown graphically in Figure The contributions of R t /D 5, S m /D 5, and HV to R 2 were 58%, 37.8%, and 4.2%, respectively. As in the previous analysis, this indicates that hardness HV has considerably less influence on the tan δ cv values than R t /D 5 and S m /D 5. ( ) Rt 3 Sm 4 tan δ cv = HV D D (R 2 =.474, N = 7) 5 5 ( ) measured tan δ cv tan δ cv from multiple linear regression Figure 3.29 Goodness of fit for multiple linear regression on Model sand tan δ cv values 59

33 3.5.5 Observations from the linear regression analyses results Based on the four linear regression analyses presented above, the following observations and comments are made: - The limited size of the samples used for the regression analyses diminishes the statistical significance of the results of the regression analyses. - All four regression equations showed a positive coefficient for the R t /D 5 term. This indicates a general trend that the interface friction coefficients (peak and residual) increase with increasing relative asperity height, R t /D 5. This concurs with the observations noted by Uesugi (1987) and Frost and Han (1999). - All four regression equations showed a negative coefficient for the relative spacing term (S m /D 5 ), indicating that the interface friction coefficients (peak and residual) decrease with increasing relative spacing, S m /D 5, which is reasonable over the range of S m /D 5 values tested. - The four regression equations showed both positive and negative coefficients for the Vickers hardness, HV, term. This, and the low contribution of this variable to the overall R 2 coefficient of the fit, indicates that hardness, HV, is not a significant predictor of the interface friction coefficients values. - The regressions show only moderate fit strengths with coefficients of determination, R 2, ranging from.47 to.89. This indicates that other factors besides R t /D 5, S m /D 5, and HV also have important influences on the values of the interface friction coefficient. - Although soil angularity could not be included in the regression analyses, it can be seen that inclusion of an angularity variable alone will not be enough to improve the strength of the fits. This can be seen in Figure 3.25 where the goodness of the fit is poor despite the fact that all points correspond to the same sand with the same angularity. Adding an angularity term to the regression equation would not increase the strength of this fit. This indicates that other 6

34 factors other than soil angularity influence the values of interface friction coefficient Influence of angularity of sand The regression analyses presented in the preceding sections did not include the influence of soil angularity. As explained earlier, this was because the angularity of the sands used in this study was only described qualitatively. The sand angularity for the Density sand was described as subrounded to rounded, and for the Model sand as subangular to angular. In general, angular sands have values of higher interface friction angle than rounded sands (Han and Frost 1999). This is consistent with results for the Model and Density sands, as listed in Table 3.8. Due to the higher contact stresses at sharp grain edges, angular soil grains are able to penetrate into the surfaces of some of the pile materials. For this reason, and because angular particles may interlock better with the pile surface roughness, interface friction angles tend to be larger for angular soil grains than for rounded grains. 3.6 SUMMARY The following laboratory activities were performed for this investigation: - Selection of sand specimens for interface testing. - Soil characterization tests to determine grain size distribution, minimum/maximum density, and specific gravity testing. - Direct shear tests on Density and Model sands to determine internal friction angles. - Selection of pile surface specimens for interface testing. - Surface topography characterization and surface hardness determination of seven pile surfaces. The seven pile surfaces included four FRP composites (commercially available from two FRP composite pile manufacturers), one recycled plastic pile, and two conventional pile materials (a prestressed concrete pile, and a steel sheet pile). 61

35 - Design and construction of a modified top half shear box to permit testing of curved pile surfaces obtained from the FRP composite tubes. - Interface shear tests for the seven piles and two sand types to determine the interface behavior and interface friction angles of these interfaces. A series of interface shear tests were performed on fourteen types of soil-pile interfaces. Tests were carried out for two types of sands: Density sand (D 5 =.5 mm, subrounded to rounded particle shape) and Model sand (D 5 =.18 mm, subangular to angular particle shape). Seven pile surfaces were tested: Lancaster FRP composite (curved), Hardcore FRP composite (curved), Hardcore FRP composite plate (flat), Hardcore FRP composite plate with bonded sand treatment (flat), PPI recycled plastic coupon (flat), prestressed concrete pile coupon (flat), and a steel sheet pile coupon (flat). The results of these tests are summarized in Appendix A. The peak and residual interface friction angles for the 14 sand-to-pile interfaces tested are summarized in Tables 3.6 through 3.8. The following points and observations can be made regarding the experimental data: - The interface friction angle values obtained for the Density sand tested against the Lancaster FRP composite pile were much lower than the values obtained for the other pile surfaces. The peak and residual interface friction angles obtained for the Lancaster FRP composite pile were 19.7 o and 16.6 o, respectively. In contrast, the interface friction angle values obtained for the Density sand tested against the other pile surface types ranged from 27.6 o to 33 o, and from 24.9 o to 27.8 o, for the peak and residual conditions, respectively. - The Lancaster FRP composite pile was measured to have the largest average mean line spacing, S m, and the second lowest maximum peak to valley height, R t (as reported in Table 3.4). These values make this pile surface the smoothest of all seven surfaces tested. - In general, the subangular to angular Model sand gives slightly higher interface friction angles than the subrounded to rounded Density sand. However, for the Lancaster FRP composite pile the interface friction angles obtained with the Model sand are much higher than the Density sand values. - In general, the interface friction angles, both peak and residual, were found to increase with increasing relative asperity height, R t /D 5. 62

36 - In general, the interface friction angles, both peak and residual, were found to increase with decreasing relative spacing, S m /D 5, which is reasonable over the range of S m /D 5 tested. - Linear regression analyses between the interface friction coefficients and the variables R t /D 5, S m /D 5, and HV, showed moderate fit strengths. The regression analyses suggest other factors besides R t /D 5, S m /D 5, and HV also have important influences on the values of interface friction coefficeint. - The bonded sand surface treatment used for the Hardcore FRP plate was successful in increasing the interface friction angle values. 63